Guide5 min read

How to see and turn off what ChatGPT remembers about you

Handy when it remembers you’re vegetarian. Less handy when it remembers everything else. Here’s how to check.

Short answer

Go to Settings → Personalization → Memory. There you see what ChatGPT remembers, can delete it, or turn Memory off. Two catches: deleting a chat doesn’t necessarily delete a memory saved from it, and turning memory off doesn’t stop training. For a question you don’t want remembered, use a temporary chat: it never creates memories.

You ask ChatGPT for a dinner idea, and it already knows you don’t eat meat. Handy. Then a thought creeps in: what else did it keep? That question about a strange mole? The long chat about your ex? Here’s how to see what ChatGPT remembers about you, and how to make it forget.

How ChatGPT remembers you

Memory isn’t one thing. ChatGPT can draw on several sources (OpenAI):

  • Saved memories. Details you asked it to remember, or that it saved by itself as useful. They’re stored separately from your chats.
  • Chat history. With Reference chat history on, ChatGPT picks up relevant things from your past conversations. There’s no separate storage limit for this.
  • The rest. Depending on your plan and country, also your custom instructions, files in your Library and connected apps like Gmail.

Since June 2026, ChatGPT is moving from a list of separate memories to a memory summary that keeps itself up to date (OpenAI release notes). OpenAI says the controls differ by plan, country and device, so your screen may look a little different from what you read here (OpenAI).

In Europe, ChatGPT asks first: in the EU, the UK, Switzerland, Norway, Iceland and Liechtenstein, remembering past chats was off by default when it arrived in 2025, and you had to switch it on yourself (OpenAI release notes).

See what it remembers

  1. Open the memory settings. Settings → Personalization → Memory. Depending on your account you’ll see a memory summary, or a list under Saved memories (OpenAI).
  2. Just ask. Type “What do you remember about me?”. OpenAI says that works. Fair warning: the summary is a “high-level view” and can leave details out (OpenAI).
  3. Check the Sources. When an answer uses something personal, Sources can appear under it, showing which past chat, memory or instruction played a part. From there you can often correct or delete it (OpenAI).

Make it forget one thing

  • Correct or delete it in the memory summary or under Saved memories. Or simply tell ChatGPT to forget it (OpenAI).
  • Careful with Don’t mention this again. It sounds like deleting, but it isn’t: ChatGPT just brings it up less. The original stays where it was (OpenAI).

Turn it all off

  • Clean slate: on the memory summary, open the ••• menu and choose Delete and turn off memory. That deletes what’s in the summary and switches Memory off. Your chats stay. Since June 2026 this works on the web; for the apps OpenAI said it was coming soon (OpenAI release notes).
  • Only past chats: turn off Reference chat history. What ChatGPT took from your past chats is then deleted from OpenAI’s systems within 30 days. Your saved memories and the chats themselves stay (OpenAI).
  • Saved memories too: turning off Reference saved memories also turns off chat history (OpenAI).

One thing to know before you switch it back on: ChatGPT may make new memories from the chats still in your history (OpenAI).

The catch: deleting a chat isn’t enough

Here’s the part almost nobody knows. Memory and chats live in different places, so cleaning up one doesn’t clean up the other (OpenAI):

  • Deleting a chat doesn’t necessarily delete a memory that was saved from it.
  • Deleting a memory doesn’t remove it from your old chats.
  • OpenAI may keep logs of deleted memories for up to 30 days, for safety and debugging.

So to really get rid of something: delete the memory, delete the chat where you said it (check your archived chats too), remove files from your Library and disconnect apps that hold it. Want everything gone, including the chats? What happens when you delete a ChatGPT chat.

Asking something personal? Melo, Kamelo’s free private chat, answers on your own computer, so your question isn’t sent anywhere. How it works

Next time: a temporary chat

For a question you don’t want remembered, start a temporary chat. It never creates or updates memories (OpenAI).

  • New since August 2026: before you start, you pick Personalized (it can use what it already knows about you) or not. By default it isn’t personalized, and you can’t change it once the chat has started (OpenAI release notes).
  • Not anonymous. It stays out of your history and isn’t used for training, but OpenAI can keep a copy for up to 30 days for safety (OpenAI).
  • Don’t save it if you want it forgotten: a saved temporary chat becomes a normal chat that memory can use (OpenAI).

Memory and training are two different switches

Turning memory off doesn’t stop training. On a personal account, OpenAI may use your chats and what ChatGPT remembers to improve its models as long as Improve the model for everyone is on. You turn that off under Settings → Data controls (OpenAI). Work accounts on Business, Enterprise or Edu aren’t used for training by default. More rights you have as a European: what you can ask OpenAI.

Where Kamelo helps

Memory keeps what you tell it. Kamelo helps before you send: when you type something personal on ChatGPT, Claude or Gemini, it suggests a temporary chat, which never creates memories, or Melo, which answers on your own computer in Chrome, so there’s nothing for any AI company to remember. Kamelo doesn’t change your message and can’t hide who you are to your AI account.